Studies in humans have largely paralleled those described above in monkeys, but are limited by inherent properties of the lab's mechanical capabilities, given the larger payload of moving humans. These studies are therefore limited to motion in the head's horizontal plane. Humans, however, provide unique opportunities to control the context in which experiments take place, including such subtleties as the tracking of imaginary targets in darkness, and to assess perceptual variables in addition to reflex behavior. We have now characterized the important features of the human LVOR and its interactions with vision during interaural oscillation across a frequency range, 0.5D4 Hz. Horizontal eye movement responses are robust, binocular, and modulated by viewing distance, as measured by vergence (angle between the eyes). The relationship between response amplitude and vergence in humans is remarkably close to that in monkeys, and shows a similar frequency-dependence. The presence of earth-fixed or headfixed targets provides a contextthat can be used to enhance or suppress the LVOR response, even in darkness (with imagined targets). However, this influence is only robust at relatively low frequencies, and follows the response dynamics typical of visual tracking. The characteristics of visual and contextual influences resemble those in the AVOR, which we have also included in detail in our subjects. We have also investigated the perception of motion in order to compare eye movement responses to psychophysical measures. A joystick task provided a measure of perceived translation. We found that the sensation of linear velocity was surprisingly transient, lasting only seconds. In contrast, when subjects were asked to respond proportionally with linear displacement in space, considerably more accurate performance was observed. These disparate results can only mean that non-vestibular cues are being used to confer improved position sense.